Your Linux Data Center Experts

beagle

Since Fedora Core 5 includes a number of mono applications, including beagle, it seemed like a good time to try it out. Read on for my impressions of this searching application…

A number of folks have already said it, but I will add my voice to the chorus of people who thought the addition of mono to Fedora Core was poorly handled. For as long as mono existed and Fedora Core existed we were told that mono had patent issues, and could not ever be included. Then right before code freeze on Fedora Core 5 test 3, it was added in, with pretty much no useful explanation. Turns out that Red Hat joined OIN, and thus got some protection from the MS patents in mono.

Since mono is now in, it made it easy to test out beagle. I had attempted to use it on my test machine running Fedora Core development after it was added, but it had a vast number of problems there (like segfaulting, not starting, etc), so with Fedora Core 5 was the first chance to see how it works.

First the good side of things: The beagle-search GUI looks great. Having it able to search and find a string in IM conversations, emails, irc logs, web pages, rss feeds and such is very nice. I have heard that this looks a lot like Apple's “spotlight” search tool. It's pretty slick and handy. The search tool is fast. There is also a command line tool to just list the path's of files that match your string. You can do some simple search modifiers like “-bob” to remove matches that have 'bob' in them.

The beagle-status command is nice to see what beagled is doing. You can easily stop beagled with a beagle-shutdown. beagled took up about 120MB of memory on my laptop, and never really increased it's footprint, which is good.

beagled notices when new web pages, emails or IM's happen and adds them right to the index then. You don't have to wait at all to search things that just happened a few minutes ago. Very handy for finding pretty recent things.

Now, all the downsides:

  • It took beagled a very very long time to index my files. I started it around noon and it was done the next morning when I got up. That's way way too slow.
  • It indexes changes as they happen, which is cool, but it causes a noticeable spike in CPU and hit on responsiveness when it happens. Ie, you get a new email and pull it in to read it, and beagled indexes it. You load a web page, and beagled indexes it. The fan on my laptop would kick up a nock on each of these actions.
  • The GUI search interface doesn't appear to have any easy way to configure what you want to open different types of documents. Probably there is some gnome gconf key, but it would be nice to set that in prefs.
  • Another thing lacking in the GUI was the ability to always show the full path to the file. Sometimes that's all I want to be able to run some command of my own devising on it. The command line client provides this however.
  • When I stopped beagled with 'beagle-shutdown' it stopped nicely. Then I started it up again and it started re-indexing everything all over. This is very unfriendly. What happened to the 12+ hours of indexing it already did?
  • Not entirely sure why this was written in mono. Couldn't a small python script and a sqlite backend provide much of this same functionality?

Because of the re-indexing and CPU fan spikes on my laptop, I have disabled beagle for now. Perhaps in a few months I will give it a try again. It looks like something that would be very nice to have, but it's not quite there in it's current state.

comments powered by Disqus

Join our other satisfied clients. Contact us today.